Regulation of the sarcoplasmic reticulum Ca(2+)-release channel requires intact annexin VI.

Annexin VI has eight highly conserved repeated domains; all other annexins have four. Díaz-Mu?oz et al. (J Biol Chem 265:15894, 1990) reported that annexin VI alters the gating properties of the ryanodine-sensitive Ca(2+)-release channel isolated from sarcoplasmic reticulum. The investigate the domain structure of rat annexin VI (67 kDa calcimedin) required for this channel regulation, various proteolytic digestions were performed. In each case, protease-resistant core polypeptides were produced. Annexin VI was digested with V8 protease and two core polypeptides were purified by Ca(2+)-dependent phospholipid binding followed by HPLC. The purified fragments were shown to be derived from the N- and C-terminal halves of annexin VI, and demonstrated differential immunoreactivity with monoclonal antibodies to rat annexin VI. While both core polypeptides retained their ability to bind phospholipids in a Ca(2+)-dependent manner, they did not regulate the sarcoplasmic reticulum Ca(2+)-dependent manner, they did not regulate the sarcoplasmic reticulum Ca(2+)-release channel as did intact annexin VI.     

The effect of metal binding on the structure of annexin V and implications for membrane binding.

The structure of annexin V, crystallised in the presence of two calcium or barium ions for each protein molecule, was solved by molecular replacement to 0.24 nm resolution. The two metal ions are found in domains I and IV, i.e. on the same side of the channel that lies in the centre of the molecule. The structures of the barium and calcium form are extremely close, the only differences localised in the metal-binding sites that lie on the surface of the molecule. The occupancies of the metal ions, however, are lower for barium than for calcium, expressing the lower affinity of the protein for the former. The packing of the annexin molecules in the crystal asymmetric unit may represent a model for the calcium driven association of membrane-bound annexins that leads to membrane fusion.     

S100A1 and S100B interactions with annexins

Members of the annexin protein family interact with members of the S100 protein family thereby forming heterotetramers in which an S100 homodimer crossbridges two copies of the pertinent annexin. Previous work has shown that S100A1 and S100B bind annexin VI in a Ca(2+)-dependent manner and that annexin VI, but not annexin V, blocks the inhibitory effect of S100A1 and S100B on intermediate filament assembly. We show here that both halves of annexin VI (i.e., the N-terminal half or annexin VI-a and the C-terminal half or annexin VI-b) bind individual S100s on unique sites and that annexin VI-b, but not annexin VI-a, blocks the ability of S100A1 and S100B to inhibit intermediate filament assembly. We also show that the C-terminal extension of S100A1 (and, by analogy, S100B), that was previously demonstrated to be critical for S100A1 and S100B binding to several target proteins including intermediate filament subunits, is not part of the S100 surface implicated in the recognition of annexin VI, annexin VI-a, or annexin VI-b. Evaluation of functional properties with a liposome stability and a calcium influx assay reveals the ability of both S100 proteins to permeabilize the membrane bilayer in a similar fashion like annexins. When tested in combinations with different annexin proteins both S100 proteins mostly lead to a decrease in the calcium influx activity although not all annexin/S100 combinations behave in the same manner. Latter observation supports the hypothesis that the S100-annexin interactions differ mechanistically depending on the particular protein partners.     

Annexin VI-mediated Loss of Spectrin during Coated Pit Budding Is Coupled to Delivery of LDL to Lysosomes

Previously we reported that annexin VI is required for the budding of clathrin-coated pits from human fibroblast plasma membranes in vitro. Here we show that annexin VI bound to the NH₂-terminal 28-kD portion of membrane spectrin is as effective as cytosolic annexin VI in supporting coated pit budding. Annexin VI–dependent budding is accompanied by the loss of ∼50% of the spectrin from the membrane and is blocked by the cysteine protease inhibitor N-acetyl-leucyl-leucyl-norleucinal (ALLN). Incubation of fibroblasts in the presence of ALLN initially blocks the uptake of low density lipoprotein (LDL), but the cells recover after 1 h and internalize LDL with normal kinetics. The LDL internalized under these conditions, however, fails to migrate to the center of the cell and is not degraded. ALLN-treated cells have twice as many coated pits and twofold more membrane clathrin, suggesting that new coated pits have assembled. Annexin VI is not required for the budding of these new coated pits and ALLN does not inhibit. Finally, microinjection of a truncated annexin VI that inhibits budding in vitro has the same effect on LDL internalization as ALLN. These findings suggest that fibroblasts are able to make at least two types of coated pits, one of which requires the annexin VI–dependent activation of a cysteine protease to disconnect the clathrin lattice from the spectrin membrane cytoskeleton during the final stages of budding.     

Endogenous cleavage of annexin I generates a truncated protein with a reduced calcium requirement for binding to neutrophil secretory vesicles and plasma membrane

We have earlier shown that an N-terminal truncated annexin I molecule, annexin I(des1-8), is generated in human neutrophils through cleavage by a membrane localized metalloprotease. The truncated protein showed differences in membrane binding among the neutrophil granule populations as compared to full-length annexin I. In this study, we investigated the cleavage capabilities of isolated neutrophil secretory vesicles and plasma membrane, and the binding of full-length annexin I and annexin I(des1-8) to these membrane fractions. Translocations were performed in vitro to secretory vesicles and plasma membrane, respectively, at different Ca(2+) concentrations. We show that the annexin I-cleaving membrane localized metalloprotease is present both in the secretory vesicles and the plasma membrane. The N-terminal truncation of annexin I gives rise to a molecule with a decreased Ca(2+) requirement for binding, both to secretory vesicles and plasma membrane. There was, thus, no difference in binding of either full-length annexin I or annexin I(des1-8) to the secretory vesicles as compared to the plasma membrane.     

Annexin VI Binds to a Synaptic Vesicle Protein, Synapsin I

Abstract: Annexin VI bound to >14 species of proteins in the whole homogenate of rat forebrain in a Ca²⁺/phosphatidylserine- or phosphatidic acid-dependent manner. When the subcellular fractions of rat forebrain were examined with a blot from a sodium dodecyl sulfate-polyacrylamide gel, each annexin VI-binding protein showed a different distribution, suggesting that annexin VI is a multifunctional protein. Of these proteins, the doublets of M_r 80,000 were enriched in the purified synaptic vesicles and were identified as synapsin I. Annexin VI bound to the head domain of synapsin I. When the binding of annexin VI to synapsin I was characterized in the native state, the affinity of the binding for Ca²⁺ ( K _Ca) was 12.6 µ M , and the affinity for annexin VI ( K _D) was ∼270 n M . Phosphorylation of synapsin I by cyclic AMP-dependent protein kinase and by Ca²⁺/calmodulin-dependent protein kinase II inhibited the annexin VI binding. The mode of the inhibition was different between the two kinases. These results indicate that annexin VI may modulate the function of synapsin I in a Ca²⁺- and phospholipid-dependent manner.     

Inhibition of EGF-dependent calcium influx by annexin VI is splice form-specific.

Annexin VI is a widely expressed calcium- and phospholipid-binding protein that lacks a clear physiological role. We now report that A431 cells expressing annexin VI are defective in their ability to sustain elevated levels of cytosolic Ca(2+) following stimulation with EGF. Other aspects of EGF receptor signaling, such as protein tyrosine phosphorylation and induction of c-fos are normal in these cells. However, EGF-mediated membrane hyperpolarization is attenuated and Ca(2+) entry abolished in cells expressing annexin VI. This effect of annexin VI was only observed for the larger of the two annexin VI splice forms, the smaller splice variant had no discernable effect on either cellular phenotype or growth rate. Inhibition of Ca(2+) influx was specific for the EGF-induced pathway; capacitative Ca(2+) influx initiated by emptying of intracellular stores was unaffected. These results provide the first evidence that the two splice forms of annexin VI have different functions.     

Membrane-bound 3D structures reveal the intrinsic flexibility of annexin VI

Several quasi-ordered arrays and three two-dimensional crystal forms of annexin VI were obtained on artificial lipid monolayers. Three-dimensional reconstructions of the crystal forms exhibit marked differences in the orientations of the two lobes, revealing flexibility of the linker between the two lobes of annexin VI. Evidence is presented that the lobes may bind the monolayer in a parallel orientation, or an antiparallel orientation, in which the second lobe is turned away from the monolayer. It is hypothesized that annexin VI may also adopt several conformations in vivo, underlying different functional roles.     

The crystal structure of calcium-bound annexin Gh1 from Gossypium hirsutum and its implications for membrane binding mechanisms of plant annexins

Plant annexins show distinct differences in comparison with their animal orthologues. In particular, the endonexin sequence, which is responsible for coordination of calcium ions in type II binding sites, is only partially conserved in plant annexins. The crystal structure of calcium-bound cotton annexin Gh1 was solved at 2.5 A resolution and shows three metal ions coordinated in the first and fourth repeat in types II and III binding sites. Although the protein has no detectable affinity for calcium in solution, in the presence of phospholipid vesicles, we determined a stoichiometry of four calcium ions per protein molecule using isothermal titration calorimetry. Further analysis of the crystal structure showed that binding of a fourth calcium ion is structurally possible in the DE loop of the first repeat. Data from this study are in agreement with the canonical membrane binding of annexins, which is facilitated by the convex surface associating with the phospholipid bilayer by a calcium bridging mechanism. In annexin Gh1, this membrane-binding state is characterized by four calcium bridges in the I/IV module of the protein and by direct interactions of several surface-exposed basic and hydrophobic residues with the phospholipid membrane. Analysis of the protein fold stability revealed that the presence of calcium lowers the thermal stability of plant annexins. Furthermore, an additional unfolding step was detected at lower temperatures, which can be explained by the anchoring of the N-terminal domain to the C-terminal core by two conserved hydrogen bonds.     

Membrane-induced folding and structure of membrane-bound annexin A1 N-terminal peptides: implications for annexin-induced membrane aggregation

Annexins constitute a family of calcium-dependent membrane-binding proteins and can be classified into two groups, depending on the length of the N-terminal domain unique for each individual annexin. The N-terminal domain of annexin A1 can adopt an α-helical conformation and has been implicated in mediating the membrane aggregation behavior of this protein. Although the calcium-independent interaction of the annexin A1 N-terminal domain has been known for some time, there was no structural information about the membrane interaction of this secondary membrane-binding site of annexin A1. This study used circular dichroism spectroscopy to show that a rat annexin A1 N-terminal peptide possesses random coil structure in aqueous buffer but an α-helical structure in the presence of small unilamellar vesicles. The binding of peptides to membranes was confirmed by surface pressure (Langmuir film balance) measurements using phosphatidylcholine/phosphatidylserine monolayers, which show a significant increase after injection of rat annexin A1 N-terminal peptides. Lamellar neutron diffraction with human and rat annexin A1 N-terminal peptides reveals an intercalation of the helical peptides with the phospholipid bilayer, with the helix axis lying parallel to the surface of membrane. Our findings confirm that phospholipid membranes assist the folding of the N-terminal peptides into α-helical structures and that this conformation enables favorable direct interactions with the membrane. The results are consistent with the hypothesis that the N-terminal domain of annexin A1 can serve as a secondary membrane binding site in the process of membrane aggregation by providing a peripheral membrane anchor.     

Nucleotide-dependent shape changes in the reverse direction motor, myosin VI

We have studied the shape of myosin VI, the actin minus-end directed motor, by negative stain and metal shadow electron microscopy. Single particle processing was used to make two-dimensional averages of the stain images, which greatly increases the clarity and allows detailed comparisons with crystal structures. A total of 169,964 particle images were obtained from two different constructs in six different states (four nucleotide states and with and without Ca²⁺). The shape of truncated apo myosin VI was very similar to the apo crystal structure, with the lever arm bent strongly backward and around the motor domain. In the full-length molecule, the C-terminal part of the tail has an additional bend taking it back across the motor domain, which may reflect a regulated state. Addition of ATP, ADP, or ATP-γS resulted in a large change, straightening the molecule from the bent shape and swinging the lever by ∼140°. Although these nucleotides would not be expected to produce the pre-powerstroke state, myosin VI in their presence was most similar to the truncated crystal structure with bound ADP-VO₄, which is thought to show the pre-powerstroke shape. The nucleotide data were therefore substantially different from expectation based on crystal structures. The full-length molecule was almost completely monomeric; only ∼1% were dimers, joined through the ends of the tail. Addition of calcium ions appeared to result in release of the second calmodulin light chain. In negatively stained molecules there was little indication of extended α-helical structure in the tail, but molecules viewed by metal shadowing had a tail ∼3× longer, 29 vs. 9 nm, part of which is likely to be a single α-helix.     

Binding to phosphatidyl serine membranes causes a conformational change in the concave face of annexin I.

Recent studies have revealed that binding of annexin I to phospholipids induces the formation of a second phospholipid binding site. It is shown that the N terminus on the concave side of membrane-bound annexin I is cleaved much faster by trypsin or cathepsin than the N terminus of the free protein. The reactivity of the unique disulfide bond located near the concave face was similarly increased by membrane binding. These results demonstrate that Ca(2+)-dependent membrane binding induces a conformational change on the concave side of the annexin I molecule and support the notion that this face of the molecule may contribute to the formation of the secondary membrane-binding site.     

LOCALIZATION OF ANNEXIN VI IN THE ADULT AND NEONATAL HEART

Annexins are a major family of intracellular Ca²⁺-binding proteins which have been implicated in a variety of cellular functions. In this paper the authors have used confocal microscopy to compare the distribution of annexin VI in vibratome sections of the rat adult left ventricle and striated muscle of the rat oesophagus. It is shown that in rat cardiac myocytes annexin VI is associated with only the sarcolemma and intercalated discs. In contrast, it is demonstrated that in rat skeletal muscle annexin VI is associated with the sarcoplasmic reticulum, in addition to the plasma membrane, suggesting that annexin VI is regulating different processes in these tissues. Also shown is that in vibratome sections of the neonatal rat left ventricle, annexin VI has a different subcellular location to that observed in the terminally differentiated adult myocyte. In these differentiating neonatal cells annexins VI is also associated with specific subcellular structures. Furthermore, using confocal microscopy of isolated myocytes the authors demonstrate that the association of annexin VI with the sarcolemma is stable even after cells are treated with the intracellular calcium chelator BAPTA-AM, to greatly deplete cytosolic calcium levels. This demonstrates that annexin VI associates tightly with the sarcolemma, and suggests that components in addition to phospholipid are involved in binding annexin VI to the membrane. These results demonstrate that the subcellular location of annexin VI is differentially regulated, and suggest that annexin VI is required for a process or processes characteristic of the sarcolemma, and of the sarcoplasmic reticulum of skeletal but not of heart muscle.     

Apo and Calcium-Bound Crystal Structures of Cytoskeletal Protein Alpha-14 Giardin (Annexin E1) from the Intestinal Protozoan Parasite Giardia lamblia

Alpha-14 giardin (annexin E1), a member of the alpha giardin family of annexins, has been shown to localize to the flagella of the intestinal protozoan parasite Giardia lamblia. Alpha giardins show a common ancestry with the annexins, a family of proteins most of which bind to phospholipids and cellular membranes in a Ca²⁺-dependent manner and are implicated in numerous membrane-related processes including cytoskeletal rearrangements and membrane organization. It has been proposed that alpha-14 giardin may play a significant role during the cytoskeletal rearrangement during differentiation of Giardia. To gain a better understanding of alpha-14 giardin's mode of action and its biological role, we have determined the three-dimensional structure of alpha-14 giardin and its phospholipid-binding properties. Here, we report the apo crystal structure of alpha-14 giardin determined in two different crystal forms as well as the Ca²⁺-bound crystal structure of alpha-14 giardin, refined to 1.9, 1.6 and 1.65 Å, respectively. Although the overall fold of alpha-14 giardin is similar to that of alpha-11 giardin, multiwavelength anomalous dispersion phasing was required to solve the alpha-14 giardin structure, indicating significant structural differences between these two members of the alpha giardin family. Unlike most annexin structures, which typically possess N-terminal domains, alpha-14 giardin is composed of only a core domain, followed by a C-terminal extension that may serve as a ligand for binding to cytoskeletal protein partners in Giardia. In the Ca²⁺-bound structure we detected five bound calcium ions, one of which is a novel, highly coordinated calcium-binding site not previously observed in annexin structures. This novel high-affinity calcium-binding site is composed of seven protein donor groups, a feature rarely observed in crystal structures. In addition, phospholipid-binding assays suggest that alpha-14 giardin exhibits calcium-dependent binding to phospholipids that coordinate cytoskeletal disassembly/assembly during differentiation of the parasite.     

S100-annexin complexes: some insights from structural studies

Several annexins have been shown to bind proteins that belong to the S100 calcium-binding protein family. The two best-characterized complexes are annexin II with p11 and annexin I with S100C, the former of which has been implicated in membrane fusion processes. We have solved the crystal structures of the complexes of p11 with annexin II N-terminus and of S100C with annexin I N-terminus. Using these structural results, as well as electron microscopy observations of liposome junctions formed in the presence of such complexes (Lambert et al., 1997 J Mol Biol 272, 42-55), we propose a computer generated model for the entire annexin II/p11 complex.     

Crystal structure of human annexin I at 2.5 A resolution.

cDNA coding for N-terminally truncated human annexin I, a member of the family of Ca(2+)-dependent phospholipid binding proteins, has been cloned and expressed in Escherichia coli. The expressed protein is biologically active, and has been purified and crystallized in space group P2(1)2(1)2(1) with cell dimensions a = 139.36 A, b = 67.50 A, and c = 42.11 A. The crystal structure has been determined by molecular replacement at 3.0 A resolution using the annexin V core structure as the search model. The average backbone deviation between these two structures is 2.34 A. The structure has been refined to an R-factor of 17.7% at 2.5 A resolution. Six calcium sites have been identified in the annexin I structure. Each is located in the loop region of the helix-loop-helix motif. Two of the six calcium sites in annexin I are not occupied in the annexin V structure. The superpositions of the corresponding loop regions in the four domains show that the calcium binding loops in annexin I can be divided into two classes: type II and type III. Both classes are different from the well-known EF-hand motif (type I).     

Annexin VI is required for budding of clathrin-coated pits.

Isolated plasma membranes attached to a solid substratum at 4 degrees C have numerous clathrin-coated pits. These pits initially are flat but become deeply invaginated after warming to 37 degrees C. The pits remain tethered to the membrane in this rounded condition unless supplied with ATP, Ca2+, and cytosol. We now show that when cytosol is treated to remove the Ca(2+)-dependent, phospholipid-binding protein annexin VI, coated pit budding no longer takes place. Addition of purified annexin VI back to the annexin VI-depleted cytosol restores budding activity to normal. Purified annexin VI alone shows only a modest budding activity, suggesting that the cytosol contains a factor(s) in addition to annexin VI that is required for full activity. Cytosol-dependent activation of annexin VI requires both ATP and Ca2+. Annexin VI appears to be not only an active component in the detachment of coated pits from the membrane but also a site for regulating the formation of coated vesicles.     

Establishment of plasma membrane polarity in mammary epithelial cells correlates with changes in prolactin trafficking and in annexin VI recruitment to membranes

Mammary epithelial cells (MEC) of lactating animals ferry large amounts of milk constituents in vesicular structures which have mostly been characterized by morphological approaches (Ollivier-Bousquet, 1998). Recently, we have shown that under conditions of lipid deprivation, perturbed prolactin traffic paralleled changes in the membrane phospholipid composition and in the cytosol versus membrane distribution of annexin VI (Ollivier-Bousquet et al., 1997). To obtain additional information on the membrane events involved in the vesicular transport of the hormone to the apical pole of the cell, we conducted a biochemical study on prolactin-containing vesicles in MEC at two different stages of differentiation. We first showed that MEC of pregnant and lactating rabbits exhibited membrane characteristics of non-polarized and polarized cells respectively, using annexin IV and the alpha-6 subunit of integrin as membrane markers. Incubation of both cell types with biotinylated prolactin for 1 h at 15 degrees C, followed by a 10-min chase at 37 degrees C revealed that prolactin transport was activated upon MEC membrane polarization. This was confirmed by subcellular fractionation of prolactin-containing vesicles on discontinuous density gradients. In non-polarized MEC, (125)I-prolactin was mainly recovered in gradient fractions enriched with endocytotic vesicles either after incubation at 15 degrees C or after a 10-min chase at 37 degrees C. In contrast, in polarized MEC, the hormone switched from endocytotic compartments to a fraction enriched in exocytotic clathrin-coated vesicles during the 10-min chase at 37 degrees C. Association of annexin VI to prolactin carriers was next studied in both non-polarized and polarized cells. Membrane compartments collected at each gradient interface were solubilized under mild conditions by Triton X-100 (TX100) and the distribution of annexin VI in TX100-insoluble and TX100-soluble fractions was analyzed by Western blotting. Upon MEC polarization, the amount of annexin VI recovered in TX100-insoluble fractions changed. Quite interestingly, it increased in a membrane fraction enriched with endocytotic clathrin-coated vesicles, suggesting that annexin VI may act as a sorting signal in prolactin transport.     

Investigating the role played by protein-lipid and protein-protein interactions in the membrane association of the p120GAP CaLB domain.

The GTPase activating protein, p120GAP, contains an amino acid sequence motif called the Ca2+-dependent lipid binding domain (CaLB) which mediates a protein-protein interaction between p120GAP and annexin VI and also binds to negatively charged phospholipids. Because membrane association of p120GAP is important for the regulation of p21 Ras activity, we have studied the roles played by Ca2+, phospholipids and annexin VI in the membrane association of p120GAP. Here we demonstrate that a truncated CaLB domain GST fusion protein (GSTGAP618-632), lacking the ability to bind to phospholipids, is able to bind to rat fibroblast membranes in a Ca2+- and concentration-dependent manner. In addition, this fusion protein also binds to annexin VI in an amino acid sequence specific but Ca2+ independent manner. Also, when bound to annexin VI in the presence of Ca2+, this fusion protein has the ability to co-bind to phosphatidylserine vesicles. Thus, annexin VI may simultaneously mediate an interaction with p120GAP and also an interaction with membrane phospholipids. This may in part explain the mechanism by which p120GAP associates with membranes in response to Ca2+ elevation and suggests the potential importance of annexin VI in the regulation of p21 Ras and the role CaLB domains may play in the specific recognition of cellular membranes.     

Mechanism of annexin I-mediated membrane aggregation

It has been proposed that annexin I has two separate interaction sites that are involved in membrane binding and aggregation, respectively. To better understand the mechanism of annexin I-mediated membrane aggregation, we investigated the properties of the inducible secondary interaction site implicated in membrane aggregation. X-ray specular reflectivity measurements showed that the thickness of annexin I layer bound to the phospholipid monolayer was 31 +/- 2 A, indicating that annexin I binds membranes as a protein monomer or monolayer. Surface plasmon resonance measurements of annexin I, V, and mutants, which allowed evaluation of membrane aggregation activity of annexin I separately from its membrane binding, revealed direct correlation between the relative membrane aggregation activity and the relative affinity of the secondary interaction site for the secondary membrane. The secondary binding was driven primarily by hydrophobic interactions, unlike calcium-mediated electrostatic primary membrane binding. Chemical cross-linking of membrane-bound annexin I showed that a significant degree of lateral association of annexin I molecules precedes its membrane aggregation. Taken together, these results support a hypothetical model of annexin I-mediated membrane aggregation, in which a laterally aggregated monolayer of membrane-bound annexin I directly interacts with a secondary membrane via its induced hydrophobic interaction site.